Very Computer

Apologies for the long post.
I am a Linux newbie.  I have RedHat 5.1, kernel v 2.0.35.  Install and
config from CD on my DELL Latitude 3500 laptop were fine.
I needed to rebuild the kernel after changing a setting from make
config. (CONFIG_NET_RADIO)

I did
make dep;make clean; make boot
and installed the new image then updated lilo.conf.
I also did make modules and make modules_install each time.
When this kernel is booted, I get the following:
Kernel Panic: VFS: Unable to mount root fs at 03:05

The GOOD kernel shows this via dmesg after reboot:
Partition check:
 hda: hda1 hda2 hda3 < hda5 hda6 >
VFS: Mounted root (ext2 filesystem) readonly.

The BAD kernel shows this instead (as it hangs <g>):
Partition check:
 hda: hda1 hda2 < hda5 hda6 hda7 hda8 hda9 > hda3 < hda10 hda11 >
[MS-DOS FS Rel. 12, FAT 16,check=n,conv=b,uid=0,gid=0,umask=022,bmap]
[me=0x8e,cs=14200,#f=184,fs=11360,fl=10560,ds=53856,de=4096,
data=54156,se=0,ts=1936097530,ls=51340,rc=0,fc=4294967295]
Transaction Block Size=512
[MS-DOS FS Rel. 12, FAT 16,check=n,conv=b,uid=0,gid=0,umask=022,bmap]
[me=0x8e,cs=14200,#f=184,fs=11360,fl=10560,ds=53856,de=4096,
data=54156,se=0,ts=1936097530,ls=51340,rc=0,fc=4294967295]
Transaction Block Size=512
Kernel Panic: VFS: Unable to mount root fs at 03:05

Partition 03:05 is the correct partition AFAIK (/dev/hda5 on the only
hard disk, and I looked at /dev to confirm the numbers)
No options were passed via LILO at boot-time.
Here are some selected values from /usr/src/linux/.config
CONFIG_BLK_DEV_FD=y
CONFIG_BLK_DEV_IDE=y
CONFIG_BLK_DEV_HD_IDE is not set
CONFIG_EXT2_FS=y
CONFIG_FAT_FS=y
CONFIG_MSDOS_FS=y
CONFIG_VFAT_FS=y

I MUST have set a config param wrong, but I have
no idea which one.

If I should include the whole config file, email

TIA,
David

--== Sent via Deja.com http://www.deja.com/ ==--
---Share what you know. Learn what you don't.---

I'll take stab at this.  I had a similar problem when I made my first
kernel under RH 6.0. I then realized I let "make xconfig" build my boot
disk driver as module instead of built into the kernel. Duhhh !  How is
the module going to load if I don't have a boot disk driver? After I
configured my boot controller, (AIC&X) in my case,to be compiled into
the kernel all was well.  BTW I had to remember to do the same for my
Ethernet card  or insmod everytime I booted.

HTH, Roger Atkinson  Unix Sys Admin

I'll take stab at this.  I had a similar problem when I made my first
kernel under RH 6.0. I then realized I let "make xconfig" build my boot
disk driver as module instead of built into the kernel. Duhhh !  How is
the module going to load if I don't have a boot disk driver? After I
configured my boot controller, (AIC&X) in my case,to be compiled into
the kernel all was well.  BTW I had to remember to do the same for my
Ethernet card  or insmod everytime I booted.

HTH, Roger Atkinson  Unix Sys Admin

Hey everyone,

Someone please take a stab at this one.  I'm having the same problem.  I've
built about 20 Kernels with different configurations and I can never seem to
see my root partition no matter what type of device or file system I try to
use.  I must be missing something.

Thanks in advance,

If you specified ext2 filesystem suport as module in your kernel config, you are
in trouble. That's like a chicken and an egg. Mount drive using module
residing on that drive.

Roman

I missed the first part of this thread, and I couldn't find it on
Dejanews, so I don't have all the information I need to comment
completely.  However, based on what I read in the above post, I have a
similar problem.  I have Slackware 3.6 with Kernel 2.0.35 installed on
an ext2 filesystem.  When I rebuild the kernel to add module support
for my ethernet card, everything seems to go OK.  When I update the
lilo.conf file and add an "append = 64MB" parm, I get the error
mentioned in the title, "unable to mount root fs."  Oddly enough, if I
specify 'linux root=/dev/hdc2' at the LILO prompt, it will work just
fine.  Also, if I change the append parameter in lilo.conf to "append
= 64MB root=/dev/hdc2" it also works just fine.  Very strange.  I've
still got some playing around to do, but I don't yet understand what's
going on.  I've marked this as a thread to watch in my newsreader to
see if anyone else can figure it out.  If I figure it out, I'll come
back with what I discover.

--
Mark Post
EDS - Operations Technical Support, Mainframe Platforms